This is a classic problem in IF -- not classic as in "hard," just classic as in "everyone asks this question at some point." Like most classic IF problems, there are a couple of examples of how to deal with this in the Recipe Book (see Ch. 10.3 "Dispensers and Supplies of Small Objects"). Both involve having an object to represent the collective group, and an object or objects to represent the takeable thing -- initially off-stage and moved into play as needed. The first one, "Extra Supplies," uses a singular object (a pen) which you could easily change to "a handful of leaves." The second example models individual slices of pizza which allows the player to have several at a time with a hard limit. (The supply is replenished with any slice the player happens to eat.)
Neither of these examples implements a truly
infinte supply of something. This would entail what is referred to as "dynamic creation of objects" -- objects which are not defined ahead of time, but created at runtime (while the player is actually playing the game). Inform does not support this as a standard feature. There is an extension -- Jesse McGrew's "Dynamic Objects," available on the website -- for this which is tres cool, but probably over - kill for what you need. Ask yourself, do you really want or need to have your players running around dropping hundreds of leaves everywhere in the game just to see how many are available? (They will, too, I promise.) In most cases, a limited supply will be enough for the player to consider the matter sufficiently implemented, without tempting them to go all leaf - crazy on your game. To paraphrase Zarf, "There's a reason why Inform has gotten along without truly dynamic objects for 15 years."
I want to add that I'm not trying to steer you away from extensions in general or that one in particular. In another thread -- not sure if it was to you or not -- I might have given that impression. Most of the extensions represent kick - ass, easy - to - use solutions to what are generally painful problems. Some newer authors are afraid to use them because (I guess) they figure, "I don't fully understand the standard package, so why try an extension?" This is absolutely not the case. Graham's philosophy with I7 is to leave most of the more complicated, unusual, or simulationy - type things to extensions, leaving the standard rules to deal with the most universal stuff. Also, for me personally, opening up extensions to see how they work has taught me a ton about the language. Definitely use them. Just not in this case. Unless you want to.
(Sorry for such a long - winded response to what was really a simple question. Five thirty am does that to me sometimes.)